How the United States wants to "develop AI on a large scale"

The Pentagon wants to generalize artificial intelligence, respecting "military ethics" and "constitutional values".

Time to Play 8 min.

Subscribers article

On the German army stand at Gamescom, the Cologne Video Game Show in August.
On the German army stand at Gamescom, the Cologne Video Game Show in August. INA FASSBENDER / AFP

In June 2018, the US Department of Defense created a Joint Center for Artificial Intelligence (AI), under the authority of General Jack Shanahan. Its strategy director, Mark Beall, was in Paris on Tuesday, 17 September, to present his joint roadmap and to establish cooperation in this area. He was received by the Minister of Armed Forces, Florence Parly, who had unveiled the French ambitions in April, and has just published a document summarizing its strategy, which has many points in common with that of his counterpart, Mark Esper.

For the French defense, the AI ​​will be first "A precious help to the decision", said Mme Parly. She will give "Unconventional intelligence capabilities", will serve "Better protect the military" and "Lighten the pain" their tasks. But also "Strengthen cyber defense". By 2025, a budget of 400 million euros is planned, plus 100 million euros annually for projects in land logistics, maintenance, battlefield mapping, intelligence or the soldier's follow-up. On this very sensitive subject, the French doctrine affirms like that of its counterpart its refusal of "killer robots".

The United States has invested billions of dollars and is already using AI, for example in the air force, for aircraft maintenance. Mark Beall answered questions from World.

What is the US defense strategy for artificial intelligence?

Our strategy is based on five pillars. The first is to design relevant systems for this technology to be used in practice. For many years, the department (of the American defense) has invested huge sums in research and development for AI. From now on, it is a question of passing from the laboratory to the real world in liaison with the industrial world.

The second axis is to develop AI on a large scale. Our pilot projects are relatively simple, but it is difficult to broadcast them in our defense organization, which is one of the largest in the world, with 4 million people. It is important to build common foundations to allow everyone to access tools: data for training, computing power, program environments. So many standardized tools, which may have been designed for the Air Force, but will be transferable in the army or navy.


Please enter your comment!
Please enter your name here